Skip to main content

Full text of "potts"

See other formats


The Framework 


Introduction to Geography, 

Climate, Topography, 

and Hydrology 

T.J. Wilkinson 

1 The Role of the Environment 

In this chapter no attempt is made to detail the physical geography and environ- 
mental context for the entire Near East, because these are treated in key refer- 
ences on regional geography (e.g., Butzer 1971, 1995; Beaumont et al. 1976; 
Brice 1978; Fisher 1978; Sanlaville 2000). Emphasis is placed upon studies con- 
ducted over the past 15 years, and this study is intended to complement earlier 
reviews of the subject (e.g., Butzer 1995; Potts 1997a; Wilkinson 2003a; Cordova 
2007); it also focuses upon topics that relate directly to human settlement, long- 
term historical trends and human responses to the environment. The narrative is 
primarily focused upon the Holocene, namely the last 10,000 years or so. The 
environmental framework extends beyond climate to include the physical geog- 
raphy of soil resources, the hydrology of ancient rivers, as well as trends in 

Debates concerning the interactions between past human societies and the 
environment continue to swing back and forth, and it is noteworthy that recent 
literature on this subject often revives earlier arguments of environmental deter- 
minism without any awareness of the original studies (Judkins et al. 2008). 
Around the turn of the 20th century, environmental determinism was a major 
intellectual movement which argued that the environment had a strong influence 
on human settlement, behavior, and even character (Huntington 1907). Ele- 
ments of this persist in recent literature arguing for the forcing effect of climate 

A Companion to the Archaeology of the Ancient Near East, First Edition. 

Edited by D.T. Potts. 

© 2012 Blackwell Publishing Ltd. Published 2012 by Blackwell Publishing Ltd. 

4 The Framework 

on the rise and decline of settlements and states (e.g., de Menocal 2001; 
Staubwasser and Weiss 2006; Kennett and Kennett 2007) and suggesting that 
societal collapse and human evolution are driven by environmental factors. A 
second group, while acknowledging the significance of the environment, take a 
more nuanced approach that sees the environment as an important factor in 
decision-making, but only one of many social, economic, and political factors 
that influence human groups (e.g., Kuzucuoglu and Marro 2007; Rosen 2007; 
Wossink 2009). Finally, a third group argues that melodramatic terms such as 
collapse are often exaggerated, whereas human societies show considerable long- 
term resilience to environmental shocks and stresses (McAnany and Yoffee 
2010a). My own position is that although the environment does have a significant 
influence on daily life and the economy, this relationship is complex, intertwined 
and sometimes indirect. 

2 Topography and the Role of Agricultural Basins 

Contrary to popular misconception, the Near East is not mainly desert but, 
rather, consists of large areas of upland plateau (e.g., Anatolia and Iran, parts of 
Yemen), extensive sedimentary basins (e.g., Mesopotamia, coastal Palestine, and 
Khuzestan), straggling uplands (e.g., Palestine, Syria) as well as forbidding moun- 
tain belts (e.g., the Zagros and Alburz mountains of Iran; the Taurus and Amanus 
in Turkey; the Hajar Mountains of Oman and the Hijaz in Saudi Arabia) (see 
Figure 1.1). Between these, and forming part of the mid-latitude arid zone, 
extend the arid deserts of Jordan, Syria, and Arabia, as well as the semi-arid steppe 
to the north. The latter were particularly important as loci of so-called secondary 
state formation in northern Iraq and the Levant. 

Early states of the 3rd and 2nd millennia BC and their characteristic settle- 
ment, the tell (Arabic meaning "mound"), were particularly concentrated in 
sedimentary basins, both the arid basins watered by the major river systems and 
those receiving sufficient rainfall for rain-fed cultivation. The latter were major 
hearths of prehistoric communities and of te//-based communities in the 4th, 3rd 
and 2nd millennia BC. Particularly significant for early states were the agricultural 
basins of the Khabur and Ghab (Syria), the Gorgan and Tehran/Qazvin plains 
(Iran), and the Amuq and the Konya plains (Turkey). Numerous archaeological 
mounds are concentrated in such plains, whereas in the neighboring uplands and 
valleys, Chalcolithic and Bronze Age settlements, although present, tend to be 
smaller and rather sparse. These basins were major centers for the cultivation 
of cereals and pulses as well as providing long-term pastoral resources for the 
large herds and flocks of Bronze Age cities. Whereas some basins extend over 
hundreds of square kilometers, others, such as the Titri§ (southern Turkey) or 
the Mamasani plain (western Iran), provided sufficient land to support only a 
few small communities perhaps dominated by a single center. Nevertheless, their 

Introduction to Geography, Climate, Topography, and Hydrology 

Figure 1.1 Map of the area discussed in the text. 

settlement history frequently extends back to the very earliest stages of 

Deep, fertile soils, characterized by mature calcareous horizons, are ideal for 
cereals and legumes, and, if fallowed, provide an in-built buffer against shortfalls 
in precipitation. Because fine-grained basin soils retain moisture better than 
shallow, frequently coarse upland soils, they can be used for long-term cultivation 
or pasturage, hence their choice for early settlements. Nevertheless, they were 
not the only land available and over the past 2,000-3,000 years settlement has 
spread considerably into the uplands of the Levant, Upper Mesopotamia, and 
parts of Turkey. This has resulted not only in the selection of more environmen- 
tally unstable locations for settlement, but in greater soil erosion. 

The agricultural basins did not exist in isolation but were linked together by 
major rivers, particularly the Tigris-Euphrates system. Consequently, areas of 
high agricultural productivity were linked to networks of communication as well 
as to other agricultural heartlands. In contrast, and particularly significant for the 
development of early states and economies, were agricultural basins that incor- 
porated routes within themselves, as was the case in southern Mesopotamia, 

6 The Framework 

where a network of rivers and artificial channels provided excellent, low-friction 
routes, ideal for the transport of bulk products. 

In addition to the agricultural basins, the climatically marginal dry steppe of 
central and northern Syria was colonized intermittently during the early Bronze 
Age, a process that left its archaeological signature in the form of large, geometric 
Kmnzhiigeln and related sites. Given their climatically marginal location, it is not 
surprising that during periods of political weakness or climatic stress such steppe 
regions were deserted or perhaps reverted to extensive pastoralist use. 

3 Rivers and River Systems 

Middle Eastern rivers and lakes provide a long-term repository for environmental 
proxy records. However, because rivers tend to remove parts of their deposits, 
the terraces that flank the rivers provide only a partial record of the history of 
deposition and erosion. Whereas traditional assessments have been constrained 
by the lack of strata suitable for direct dating (Geyer and Monchambert 2003), 
new radiometric dating techniques on basalt flows interleaved with terrace gravels 
are pushing the chronology of the Euphrates and Orontes Rivers back to some 
2-3 million years before present (BP), indicating that these river valleys are sig- 
nificantly older than originally thought (Demir et al. 2007). 

The fertile riverine terraces provided extensive agricultural resources close to 
water and communications and therefore attracted large-scale settlement. 
However, rivers also provided recognizable boundaries and thus functioned as 
frontiers, as illustrated by the Syrian Euphrates in the Neo-Assyrian and Roman 
periods. Consequently, river valleys should not simply be seen as loci of seden- 
tary communities. This tension between a valley's varied roles - in agricultural 
settlement, as the location of a boundary and as a corridor of movement - 
perhaps explains why some riverine corridors show oscillating patterns of human 
settlement, in one period acting as attractors, and in another being deserted. 
Hence, the Euphrates River in Syria and Turkey was a major locus of Early 
Bronze Age settlement during the 3rd millennium BC, with settlements fre- 
quently located on opposite banks of the river, presumably at crossing points 
(Wilkinson 1990). Equally though, river valleys such as that of the Sajur in 
northern Syria, were sparsely settled for much of the 4th and 3rd millennia BC, 
despite their apparent fertility, and only densely settled from the Iron Age 
onwards. Such examples are instructive in demonstrating that the environment 
should not be viewed as a single, monolithic resource, but as a set of potential 
resources affording opportunities that could be taken up at different times in 
different ways. 

The Tigris -Euphrates river basin not only formed one hearth for the origins 
of agriculture, but was also a key for the urban revolution of the Early Bronze 
Age. As noted by Algaze, it formed "an enormous dendritic transportation 

Introduction to Geography, Climate, Topography, and Hydrology 7 

system" (2001: 204a) available to float the products of Anatolia, Iran, and Syria 
down to the Mesopotamian lowlands. 

Many Middle Eastern rivers exhibit ephemeral flow and vary considerably in 
their geometry. They include braided gravel-bed channels, meandering single- 
thread channels within silt-clay flood plains, as well as rather straight channels 
with very gentle gradients. The last named class of channel is particularly common 
in the Mesopotamian lowlands, where it forms part of the anastomosing system 
of channels that thread back and forth across the alluvial plains. 

At first glance, the southern Mesopotamian plains appear flat, bleak, and arid, 
presenting an austere vista broken only by linear palm gardens, upstanding canal 
banks, archaeological mounds, and occasional dune fields. Further scrutiny reveals 
occasional low alluvial levees which have built up over the millennia as a result 
of the preferential deposition of sand and silt along rivers and other channels. 
These ancient levees form substantial landscape features that stand up to five 
meters above the surrounding flood basins and may be five kilometers wide. 

The Mesopotamian alluvial plains developed over the Quaternary period of 
geological history, with much fine clay, silt and sand being deposited over the 
past 10,000 years (i.e., the Holocene; Buringh 1960: 162). Nevertheless, parts 
of the plain date back to the Pleistocene period, and occasional islands of Pleis- 
tocene deposits have been recognized in the northern plains near Sippar, as well 
as further south near Tell Oueli and Larsa (Geyer and Sanlaville 1996), possibly 
talcing the form of "turtlebacks" near Uruk (Pournelle 2003: 6 n3). Rather than 
forming a stack of fine-grained sediments, they accumulated episodically with 
periods of sedimentation punctuated by gaps or hiatuses. Sediments preferentially 
accumulated along river channels so that the channel became raised and the flood 
plain aggraded. Floods accumulate in basins that become sumps for clay deposi- 
tion as well as fresh or brackish marshes. If a river shifted its channel, or the 
protective cover of vegetation was lost (because of lack of soil moisture), the soils 
would dry out and become vulnerable to wind erosion, with the eroded topsoil 
being incorporated into dunes of sand and silt-clay aggregates. As a result of the 
two contrasting processes of alluvial accumulation and aeolian degradation, parts 
of the archaeological record may be buried whereas, nearby, the very earliest 
archaeological sites are revealed (Wilkinson 2003a). Whereas one area may reveal 
a wide range of settlement back to the Ubaid period (6th/5th millennia BC), 
other areas are confined to more recent settlements of Parthian, Sasanian, and 
Islamic date. Although not as complex as the Mesopotamian lowlands, the Khuz- 
estan plain of southwestern Iran also shows critical lateral variations in sedimenta- 
tion that result in differential preservation of prehistoric sites (Kouchoukos and 
Hole 2003). 

River channels and associated levees provided a focus for settlements and their 
associated palm gardens, and a source for irrigation water which often flowed 
down the levee slopes along a branching or trellised system of canals. The 
archaeological surveys of Robert McCormick Adams (1981) demonstrate how 

8 The Framework 

settlements from the 4th millennium BC onward were preferentially sited along 
levees, and because many levees are associated with sites dating from prehistoric 
to Islamic times, it is clear that the associated channels formed long-term features 
of the landscape. 

Whereas much has been made of the role of changing climate in the history 
of human settlement, rather less has been said about abrupt shifts in river chan- 
nels known as avulsions (Gibson 1973). If the abrupt shift of a river channel 
occurred from the levee crest to a more convenient course, a long was often 
obligated to redirect flow back into the vacated channel or to dig new canals in 
order to guarantee the continued supply of water for irrigation (Cole and Gasche 
1998: 14; cf. Potts 1997: 24-5a). Such abrupt shifts have been recorded in 
Mesopotamia in recent times, and hinted at in antiquity as, for example, when 
Sin-iddinam, long of Larsa (c. 1849-1843 BC), stated that, "in order to provide 
sweet water for the cities of my country . . . [An and Enlil] commissioned me to 
excavate the Tigris [and] to restore it [to its original bed]" (Frayne 1990: 
158-60). One of the roles of the king was to ensure that favorable conditions 
existed, and if a river shifted its course, it was necessary to try to shift it back 

Geomorphological studies suggest that avulsions were probably a significant 
factor in antiquity as well. They have been documented at Rut on the Tigris 
(Verhoeven 1998; Hritz 2004; Heyvaert and Baeteman 2008), as well as near 
Samarra (Adams 1965), and the overall branching tendency of the lower Tigris 
and Euphrates rivers suggests that avulsion may have been the dominant process 
in the development of the anastomosing Mesopotamian rivers (Wilkinson 2003a: 
83-9). Others have argued that it was even a significant factor in the cyclical 
changes of Mesopotamian civilizations (Gibson 1973; Morozova 2004). The 
branching channels of the Euphrates near Tell ed-Der and Sippar (Iraq) provide 
a good example of a node of avulsion, off which the Irnina, Purattum (main 
branch), and Purattum (Kish branch) of the Euphrates (Purattum) branched 
during the 3rd, 2nd and 1st millennia BC (Cole and Gasche 1998; Heyvaert and 
Baeteman 2008). Dealing with the physical problems created by such channel 
shifts is also reflected in ritual and religious practice, and it is probably no coin- 
cidence that texts from the Sippar region, an early cult center of the river god, 
contain an unusually high number of references to it (Woods 2005). 

Avulsions and branching river channels are not limited to the Mesopotamian 
plains, however. Major channel shifts have also been recorded in the Khuzestan 
(Moghaddam and Miri 2007) and Gorgan plains (Omrani Rekavandi et al. 2007, 
2008) in Iran. 

An additional by-product of human activity is salinization, a significant problem 
for agricultural production in southern Mesopotamia, parts of southwestern and 
northeastern Iran, and the Konya basin in Turkey. Saline soils contain soluble 
salts in such quantities that they interfere with the growth of most crop plants, 
and salinization is the principal soil-forming process in central and southern Iraq 

Introduction to Geography, Climate, Topography, and Hydrology 9 

(Buringh 1960: 83). In southern Iraq the high salt content of river waters, high 
levels of evaporation, low soil permeability, raised water tables, low gradients, 
and application of too much irrigation water result in poor drainage and a pro- 
gressive build up of salts which reduces crop yields. Fortunately, salt accumula- 
tions are mitigated by traditional crop husbandry. By fallowing the land every 
second year, vegetation growth is able to lower the water table so that accumu- 
lated salts are leached away by the next flush of irrigation waters, thereby reduc- 
ing, but not eliminating, salinization. The old theory that salinization in southern 
Mesopotamia led to a shift from the cultivation of wheat to salt-tolerant barley, 
reduced cereal production, and perhaps led to the abandonment of cities has 
been widely criticized (Postgate 1992: 180-1; Powell 1985). Nevertheless, soil 
scientists have demonstrated that salinization is endemic in southern Mesopota- 
mia and its eradication is difficult (Buringh 1960). Consequently, at a local level 
salinization must have posed significant problems for ancient agriculturalists. 

Although the major rivers of the Middle East formed an enduring part of eve- 
ryday life, those ancient communities that relied upon them for their livelihood 
also needed to deal with the instabilities inevitably associated with them. In addi- 
tion to avulsion and salinization, these included the 3rd millennium BC mega 
floods characteristic of the Euphrates river (Oguchi and Oguchi 1998) and river- 
ine incision that left flood plains high and dry, thereby making floodwater farming 
impossible (Rosen 2007: 177). Humans were agents in the creation of instabilities 
by narrowing rivers, thereby raising water levels and increasing the risk of flood- 
ing, and encouraging rivers to stay in their present channels. Levees were raised 
so that if a channel break did occur, the resultant flooding could be devastating. 
River banks were weakened by digging canal off-takes and over-irrigation encour- 
aged the elevation of water tables and associated salinization. Finally, clearing 
woodland increased run-off to rivers, increasing chances of flooding. 

However, humans could mitigate natural processes by irrigating to ensure 
against drought years, by employing fallow to accumulate soil moisture as insur- 
ance against seasons of low rainfall, by using fallow to lower the water table and 
limit the affects of waterlogging and salinization, and by investing in check dams 
across seasonal wadis to limit incision. The overzealous management of natural 
processes could result in problems which would either impact the communities 
themselves (e.g., salinization), or affect downstream communities, as occurred 
with water abstraction, avulsion, or deforestation near the headwaters of major 
rivers. Such problems may have been exacerbated by increasing population, over- 
ambitious kings or the use of natural resources as a weapon in times of conflict 
(Cole 1994). Ignoring land management practices such as fallowing may have 
pushed a theoretically sustainable system of agriculture toward a threshold where 
either two or three dry years (in the rain-fed north) or the annual application of 
irrigation water (in the irrigated south) would have precipitated agricultural fail- 
ures that would not have happened if traditional, sustainable practices had been 


The Framework 

4 Climate and Climatic Change 

The pioneering studies of Lamb (1977) and Butzer (1958) assembled a wide 
range of evidence indicating that over the last 10,000 years the climate has varied 
through cycles of increased moisture, heat, and storminess. Unfortunately, since 
the late 1960s a new orthodoxy has suggested that, over the past 6,000 years, 
and for perhaps virtually the entire Holocene, the climate has been essentially 
stable (Raikes 1967). This stable Holocene scenario was reinforced by initial 
interpretations of the Greenland ice cores, which indicated a much less tempestu- 
ous Holocene than had prevailed during the preceding glacial phases (Anderson, 
Goudie, & Parker 2007: 148-149). The earlier studies have now been vindicated, 
as investigations away from the polar regions have demonstrated the true variabil- 
ity of the Holocene climate. In addition to minor fluctuations, six episodes of 
"rapid climate change" have been identified over the last 8,200 years (Anderson, 
Goudie, & Parker 2007: Fig. 1.2). However, the definition of what constitutes 
a climatic "spike" is subjective, and their effect on past human activity in the 
Near East remains hotly contested (Weiss et al. 1993; Butzer 1997; Zetder 2003; 
Kuzucuoglu and Marro 2007). 


Period of BC 

the later 



Early Bronze 
Age states 

Uruk- Northern Pre-Pottery 

LChalc Ubaid Halaf Neolithic 
5000 6000 7000 8000 9000 10,000 11,000 12,000 13,000 14,000 15,000 BP 

-i 1 1 1 1 1 1 \ wetter 


3000 4000 5000 6000 7000 8000 9000 10,000 11,000 12,000 


Figure 1.2 Reconstruction of wetter and drier periods in the climate record of Lake 
Van (Turkey) and Qunf Cave (Oman). 

Introduction to Geography, Climate, Topography, and Hydrology 11 

Thanks to a substantial increase in the budget for environmental research 
during the past 30 years, as well as an increase in the number and sensitivity of 
techniques of analysis, the amount and geographical distribution of climate proxy 
records has increased tremendously. 

Butzer (1978: 6-11) defined three broad paleoenvironmental zones within 
the Middle East: 

• a northern highland perimeter consisting of Anatolia, Armenia, Kurdistan and 
Iran which receives much of its rainfall in winter; 

• the hills and plains of the Fertile Crescent including the Levant and 

• the desert belt of northeastern Africa and Arabia, the southern parts of which 
receive most rainfall (when it occurs) as summer monsoonal rain. 

The Arabian desert, an ancient feature extending back millions of years, falls 
within the warm desert zone of the northern hemisphere (Anderson, Goudie, & 
Parker 2007: 121). As a result of differences in the earth's orbital parameters and 
the amount of solar radiation received, the Sahara, Arabian, and North Indian 
deserts received significantly more rainfall between about 10,000 and 6000 BP 
than they do today. Warmer ocean temperatures in the tropics were associated 
with a strengthened monsoonal system and greater atmospheric moisture (Kut- 
zbach and Liu 1997; Brooks 2006). On the other hand, the northern highland 
perimeter and Fertile Crescent steppe receive significant precipitation in winter 
from depressions moving along the path of westerly circulation. It is noteworthy 
that the flow of the Tigris-Euphrates and tributaries, which fall within the first 
two regions, shows a partial correlation with the so-called North Atlantic Oscil- 
lation which governs the path of mid-latitude storm tracks and precipitation 
within the Mediterranean region (Cullen et al. 2000). 

However, climate changes in the region are not simply a result of variations 
in the westerlies to the north versus the monsoonal system to the south. Recent 
high-resolution studies of lake sediments from Nar Golii (central Turkey) show 
links to both the north Atlantic and Indian Ocean monsoons (Jones et al. 2006: 
364; Cordova 2007: 130) so that increased summer aridity in central Turkey 
correlates with periods of enhanced Indian monsoonal rainfall (Jones et al. 2006: 
363). In other words, drier conditions in the eastern Mediterranean can be linked 
to a regionally complex monsoon evolution which itself is related to patterns of 
upper atmosphere airflow in the sub-tropics (Staubwasser and Weiss 2006). 

Because we lack tools for directly measuring climatic warmth or moisture, it 
is necessary to employ so-called proxy records - i.e., indirect indicators such as 
pollen, carbon/oxygen isotopes, and micro-organisms in lake sediments that vary 
according to climate. The most common proxy tools used include lake sediments, 
soils, and caves for carbon and oxygen isotope studies; lakes and marshes 
for palynology; and rivers for records of long-term hydrology. In addition, the 

12 The Framework 

surrounding seas and oceans provide long-term proxy records, especially for 
ocean temperature, patterns of global circulation, and continental river dis- 
charges. No technique provides a simple and straightforward record of the past 
environment because each record will have accumulated under different circum- 
stances, so that breaks in the sequence may occur. For example, lake records have 
the great advantage of (frequently) providing continuous records of sedimenta- 
tion from which the proxy indicators can be sampled. By contrast, rivers and 
wadis are characterized by spatial variability and a tendency for the river to remove 
earlier deposits creating gaps and discontinuities. Nevertheless, some lake sedi- 
ments are remarkably discontinuous, Lake Konya being a particularly good 
example (Fontugne et al. 1999); others show considerable variations in the rate 
of sedimentation, which needs to be allowed for when computing the proxy 
records. Moreover, indicators such as carbon and oxygen isotopes are themselves 
formed within a complex geochemical environment, so that each record must be 
interpreted on its own merits (Jones and Roberts 2008: 37). Although precipita- 
tion and evaporation are generally seen as driving the hydrological context of 
these isotopes, the source area of the rainfall and the temperature of the lake 
water itself are significant. Consequently, identifications of wetter and drier 
phases in climatic records must always be regarded cautiously (Jones and Roberts 
2008; Develle et al. 2010). 

Pollen provides a record of past vegetation, but accounts vary according to 
pollen productivity as well as the size and catchment of the lake or marsh. Frus- 
tratingly, most lakes are located within the moister parts of the Middle East, 
hence many early sites in north Syria and Mesopotamia are far away from the 
nearest proxy climate record. However, the mere existence of lakes in areas where 
today the environment is arid, such as a Green Arabia that was rendered more 
verdant as a result of strengthened monsoon circulation (McClure 1978; Lezine 
et al. 2007; Parker et al. 2004), provides an indication of periods of increased 
atmospheric moisture. 

Bearing the above caveats in mind, the following places provide valuable 
records of climate change for the Middle East (see Figure 1.1 above): 

• Lake Van (Lemcke and Sturm 1997; Wick et al. 2003) and Eski Acigol 
(Roberts, Reed et al. 2001) in Turkey; 

• the Beqaa valley and mountains of Lebanon (Develle et al. 2010; Verheyden 
et al. 2008; Hajar et al. 2010); 

• the Ghab valley in Syria (Yasuda et al. 2000); 

• the Dead Sea (Neumann et al. 2010; Migowski et al. 2006; Leroy 2010) and 
Soreq Cave (Bar-Matthews et al. 1997; Bar Matthews and Ayalon 2011) in 
Israel, Palestine, and Jordan; 

• Qunf Cave in Oman (Fleitmann et al. 2003); 

• Hawa and Dhamar lakes in Yemen (Lezine et al. 2007; Davies 2006); 

• Awafi lake in the United Arab Emirates (Parker et al. 2004); 

Introduction to Geography, Climate, Topography, and Hydrology 13 

• Lakes Zeribar, Miribad, and Almalou in Iran (Snyder et al. 2001; Djamali, 
de Beaulieu, Andrieu-Ponel et al. 2009: 1372). 

Of these, Lake Van provides a valuable record of westerly circulation within 
Anatolia, and Qunf Cave acts as a sensitive indicator of monsoonal circulation in 
southern Arabia (see Figure 1.2). The Van record, which derives from annual 
laminations corresponding to seasonal deposition of silt, can be summarized as 
follows (Lemcke and Sturm 1997; Wick et al. 2003; Kuzucuoglu 2007: 468): 

• From approximately 5500 BC (later Neolithic) until around 3050 BC (begin- 
ning of the Early Bronze Age) the climate was moister than that of today. 

• There followed a transition from 3050 to 2050 BC with increasing aridity 
interrupted by drier phases at 2400 and 2150 BC and generally drier condi- 
tions in the final quarter of the millennium. 

• After 2050 BC (from the Middle Bronze Age), conditions remained drier, 
becoming similar to the present day around 1 BC. 

Despite regional variations, there is some consensus on the broad trends in 
climate change for the northern Near East (Robinson et al. 2006: 1535; but see 
Roberts et al. 2011 for a variant chronology): 

• The Late Glacial Maximum (23,000-19,000 BP) was cooler and more arid 
than the present day, but Lake Lisan, the predecessor of the Dead Sea, was 
high due to an excess of water supply from precipitation, runoff and ground- 
water discharge over evaporation. 

• Localized wanning trends followed, but the cold episode of the Younger 
Dryas (12,700-11,500 BP) ushered in a millennium or so of cold, arid 
conditions. Lake Lisan dropped significantly and large deposits of salt 

• The Early Holocene, from c.9500 to 7000 BP, was the wettest phase in the 
last 25,000 years in the Levant and eastern Mediterranean, and the margin 
of the Negev Desert migrated significantly further south; the level of the Dead 
Sea was relatively high. 

• A brief moist episode followed around 5000 BP, after which the climate 
became significantly drier; there is consensus that the climate of the northern 
Near East became drier during and especially after the Early Bronze Age 
(c.4200 or 4000 BP) but the details vary depending upon the record. 

The Dead Sea and Soreq Cave provide excellent, long-term records that are 
complemented by high-resolution records of tree rings from sub-fossil tamarix 
trees within Sedom Cave near the Dead Sea. Carbon and nitrogen isotopes, 
extracted from the cellulose of tamarix tree rings, show a gradual but fluctuating 

14 The Framework 

drying from 2265 to 1930 BC and maximum aridity between 2000 and 1900 BC 
(Frumkin 2009: Fig. 7). This relatively precise record, which corresponds to 
a significant drop in the level of the Dead Sea (Frumkin 2009: 326), provides 
an indicator of annual drought that would have been experienced by local 

By the Roman, Arsacid, and Sasanian periods (c. 247/2 38 BC to 651 AD) 
conditions fluctuated around those of today, and again, detailed records provide 
evidence of specific fluctuations. For example, a comparison of Mediterranean 
Sea cores with the Soreq Cave isotope record suggests humid phases at c.1200 
BC, 700 AD, and 1300 AD, and drier phases at 100 BC, 1100 AD, and 1700 AD 
(Bar-Matthews et al. 1997; Schilman et al 2002; Rosen 2007: 90). When com- 
pared with high-resolution proxy records from Nar Golu in central Turkey, there 
is some overlap, specifically in humid intervals at 560-750 AD and 1000-1350 
AD and a dry phase from 1400 to 1950 AD (Jones et al. 2006). Moreover, because 
the warm, humid phase of 1000-1350 AD and the dry, cool phase after 1400 AD 
fall within the medieval warm phase and the Little Ice Age, as known from his- 
torical sources, there is good reason to view these as credible. 

Although such comparisons provide grounds for optimism, when these records 
are compared with the environmental crises noted by Michael the Syrian, the 
correspondence of climatic peaks and troughs are difficult to reconcile. This is 
partly because the textual records mention not only occasional droughts, but also 
frequent swarms of locusts and episodes of bubonic plague as well as cold winters, 
the latter often having a devastating effect on flocks and herds (Widell 2007). 
Although even fine-grained climate proxy records are difficult to harmonize with 
historical records, in part because very different types of record are being com- 
pared, they take us significantly beyond the previous generation of coarse-grained 
proxy records. 

Whereas most proxy records from lakes and marshes are located far from major 
centers of Neolithic and Bronze Age settlement, calcareous accretions in archaeo- 
logical sediments, although lacking the fine resolution of, for example the Van 
sequence, have the advantage of providing a proxy record from within core areas 
of settlement. Calcareous coatings on stones at Neolithic Gobekli Tepe provide 
a proxy record for upper Mesopotamia, a core region for the domestication of 
plants and animals as well as early state development (Pustovoytov et al. 2007). 
The two main periods of coating formation date to the Early (10,000-6000 BP) 
and Mid-Holocene (6000-4000 BP). Fluctuations of the oxygen and carbon 
isotopes suggest that whereas the Early Holocene coatings accumulated within 
a climate of rising temperature, in the Middle-Holocene both rate of growth and 
isotopic signatures of coatings imply maximum humidity between c.6000 and 
4000 BP (Pustovoytov et al. 2007: 325). Although there is some discrepancy in 
the timing of the Middle-Holocene moist phases in the Gobekli and Van data, 
both agree that after 4000 BP (i.e. around 2000 BC) the climate of the Fertile 
Crescent became significantly drier. 

Introduction to Geography, Climate, Topography, and Hydrology 15 

Climate proxy records from the desert belts include that from Qunf Cave in 
southern Oman (Fleitmann et al 2003) where oxygen isotopes (8 18 O) from the 
growth rings of a stalagmite which grew from 10,300 to 2700 and from 1400 
to 400 BP have been documented with a time resolution of 4-5 years. The iso- 
topic record showed four main features: 

• a rapid increase in monsoonal precipitation between 10,300 and 9600 BP; 

• high monsoonal precipitation between 9600 and 5500 BP; 

• a long-term decline in precipitation from 8000 to 2700 BP; and 

• cessation of stalagmite formation at 2700 BP and regrowth from 1400 to 
400 BP. 

The latter growth phases are within the 8 18 O range of modern stalagmites and 
are therefore equivalent to the present-day climate. Although records from paleo- 
lakes and oceanic sediments do not entirely agree with the Qunf Cave evidence, 
the general pattern of a moister Early and mid-Holocene is clear. A variant date 
comes from al-Hawa (Yemeni desert), where the onset of the wet Early Holocene 
began at c. 12,000 BP (Lezine et al. 2007: 247), and the lakes in the Yemen 
highlands where it began at c. 11,000 BP (Davies 2006; Parker 2009). In con- 
trast, at Awafi in the UAE well to the north of Qunf Cave, the lake sequence 
started significantly later, at c.8500 BP, the onset of wet conditions there being 
related to the northward migration of the Inter-Tropical Convergence Zone 
(Parker et al. 2004). 

Despite their location in different latitudinal zones, the trend of the isotope 
records from Qunf and Soreq is broadly similar but with differences in detail, 
especially during the later periods. The challenge now is to interpret the settle- 
ment and historical records in light of these records in a way that allows for the 
often idiosyncratic behavior of human populations. 

5 Vegetation 

The vegetation record in the Middle East derives mainly from carbonized plant 
remains and palynology. The representativeness of plant remains on archaeologi- 
cal sites is biased by their deliberate selection by humans. Thus, they do not 
necessarily reflect the totality of plant life in a given locale. On the other hand, 
pollen sequences are usually limited to those species present around lakes and 
marshes, which are not always in the vicinity of major archaeological sites (Miller 
1998). Despite these biases, thanks to pioneering research by Van Zeist and 
Bottema (1991), together with more recent palynology, we now have a broad 
idea of the pattern of vegetation within the northern mountain fringe, and Fertile 
Crescent and Iran. According to Roberts and colleagues, the range of tree cover 
around the Mediterranean basin contracted during glacial phases and the Late 

16 The Framework 

Glacial Maximum when much of the region was dominated by herb species such 
as Artemisia and various chenopods (Roberts, Meadows, & Dodson 2001: 632). 
Tree species such as European oak, Pistacia, and olive survived only within 
limited areas such as the southern Levant. 

During the Early and Middle-Holocene the herbaceous Late Glacial steppe 
was replaced by sub-humid forest and, depending upon location, by broad- 
leaved, deciduous trees. However, forest re-advance was slow during the Holocene 
and pollen sequences from eastern Turkey and western Iran suggest that tree 
pollen approached modern values only by 7000-5000 BP, perhaps because its 
spread was inhibited by the drier, Early Holocene climate (Roberts, Meadows, 
& Dodson 2001: 632). However, in parts of the southern Levant where decidu- 
ous rather than evergreen oaks prevailed, the typical winter rainfall regime may 
have been replaced by summer rain (Roberts, Meadows, & Dodson 2001: 632). 

The archaeobotanists Gordon Hillman and colleagues have employed rainfall 
and temperature levels, together with the distribution of present vegetation, to 
reconstruct the potential vegetation cover of semi-arid northern Syria (Hillman 
2000). Their maps suggest that park woodland and woodland steppe extended 
into what is today the denuded and treeless north Syrian steppe. Such reconstruc- 
tions suggest that massive loss of natural vegetation occurred during the Holocene 
and that grasslands and park woodlands were replaced by an essentially desertified 
steppe. This picture is now supplemented by carbonized wood (charcoal) analyses 
which reveal a much richer vegetation in the Bronze Age than today: riverine 
vegetation was more diverse, and oak park woodland occupied many parts of the 
northern Fertile Crescent, roughly along the line of the Turkish/Syrian border 
as well as in northern Syria (Deckers and Pessin 2010: 220-1). "Massive degrada- 
tion including deforestation," as a result of both human and climatic impacts, 
the latter indicated by a northward shift of the Pistacia- woodland steppe, has 
caused the present virtually treeless landscape (Deckers and Pessin 2010: 225). 
In the Khabur basin, which is today devoid of trees, oak park woodland was still 
present as late as the 3rd century AD (Deckers and Rehl 2007: 347). 

Historical sources also provide a strong case for human removal of woodland. 
For Mount Lebanon, textual records demonstrate how the original cedar and 
coniferous woodland was probably destroyed in part as a result of expeditions of 
the Assyrians and the Akkadians before them (Rowton 1967; Mikesell 1969). 

Particularly convincing evidence of the loss of woodland and the appearance 
of orchard cultivation derives from small lakes with local catchments. The pollen 
diagram from the small crater lake of Birkat Ram, in the occupied Golan heights 
(Schwab et al. 2004), which covers the last 6,500 years, shows the first significant 
human impact on vegetation in the Chalcolithic with a decline in deciduous oak 
(see Figure 1.3). During the Middle and Late Bronze Ages and Early Iron Age, 
when the area was sparsely settled, woodland regenerated. This was followed by 
a sharp drop in deciduous oak in the mid- 1st millennium BC, or slightly later, 
perhaps as a result of ore -smelting and land clearance associated with settlement 

Introduction to Geography, Climate, Topography, and Hydrology 


years cal.BC/AD 

1185-1285 AD 


670-870 AD 
780-990 AD 
690-900 AD 
980-1190 AD 
170-390 AD 
130-350 AD 



970-790 BC 

830-660 BC 

1520-1390 BC 



2880-2580 BC 

4350-4160 BC 

m i 

4950-4720 BC 

) i 

5060-4840 BC 

d 1 

Total tree Olive 
pollen (Olea) 

Birket Ram 



- 5000 





Varve years 
before 1990 ad 

Lake Van 

(Total tree pollen) 

Figure 1.3 Pollen diagrams from Birkat Ram (Golan Heights) and Lake Van (Turkey). 

and agriculture (Dar 1993). The rising sedimentation rate at this time probably 
resulted from higher soil erosion associated with increased settlement. Human 
impact on vegetation structure is evident from the curves of olive and grape pollen 
which expanded rapidly during the Hellenistic, Roman, and Byzantine periods, 
after which olive production collapsed. There was then regrowth of evergreen 
oak, a characteristic feature of other pollen charts in the region (Baruch 1986). 
Overall, the woodland decline and associated growth of olive and grape between 
the 3rd/4th centuries BC and the 7th century AD appear to reflect the character- 
istic settlement pattern of the Levant when many upland or formerly marginal 
areas were settled and prospered, in part as a result of increased trade in olive oil 
and wine (Wilkinson 2003a: 128-50). This pattern is associated with the devel- 
opment of maquis (evergreen shrubs) and garrigue (perennial scrub) vegetation 
on many uplands as well as desertification in drier areas (Wilkinson 2003a: 150). 
Despite the long settlement record of the Near East, human impacts on veg- 
etation are less clear than might be expected, in part because Neolithic impacts 
often occurred before vegetation had stabilized following the last glaciation 
(Roberts 1998: 188). Nevertheless, the removal of oak woodland is indicated in 
the Ghab valley (Syria) during the Neolithic (Yasuda 2000), whereas clearance 
shows up in pollen diagrams by the 3rd millennium BC (Roberts, Meadows, & 
Dodson 2001: 634). A conspicuous clearance horizon is the so-called Beysehir 

18 The Framework 

phase, which appears in southwestern Turkey c. 3200-1250 BP (1250 BC and 
800 AD) (Bottema and Woldring 1990; Eastwood et al. 1998; Roberts 1998: 
188). In the Berakat basin, also in southwestern Turkey, this phase is dated 
c. 2230-1550 BP, during a period of local Hellenistic and Roman settlement 
(Kaniewski, Paulissen, De Laet et al. 2008), and is associated with increased 
burning of woodland, as well as polyculture and significantly increased soil 
erosion, presumably resulting from forest clearance and agriculture (Kaniewski, 
Paulissen, De Laet et al. 2008: 234). 

A similar phase of vegetation change recorded in the small Lake Almalou at 
2500 meters in northwestern Iran shows high Cerealia-type pollen correspond- 
ing presumably to cultivation and mobile pastoralist activity. These episodes more 
closely relate to historical rather than climatic phases, under the Achaemenid 
empire (539-330 BC) and during a phase of large-scale fruit tree cultivation in 
the Sasanian period (c. 240-640 AD) (Djamali, de Beaulier, Andrieu-Ponel et al. 
2009: 1372). 

The propagation of tree crops was not restricted to the later empires because 
Sumerian texts from the Ur III period (2100-2000 BC) demonstrate that around 
Garshana and Zabala in southern Mesopotamia, in addition to the expected date 
palms and tamarisks, some 25,000 pine trees were recorded, as well as olives, 
apples, and other fruit trees (Heimpel n.d.). Together with others such as box, 
these trees must have been introduced from the surrounding mountains or more 
temperate regions, much as the later Neo-Assyrian kings introduced foreign 
plants to their parks and gardens. 

Cultivated food crops reflect the prevailing environment and give us an idea 
of how societies coped with climatic fluctuations. Land use and crop yields are 
influenced by rainfall, fallowing practices, and soil quality, and if the last two 
factors remain constant in a given region, changes in drought-tolerant crops can 
be compared to climate proxy records. In order to determine relationships 
between crop plants and Late Holocene aridity, Rehl (2008) used carbonized 
plant remains to infer how the proportion of crop types changed according to 
rainfall as indicated by climate proxy curves. Stable carbon isotopes extracted 
from barley from seven Bronze Age sites provided a proxy for ancient moisture, 
with low A 13 C (in semi-arid environments) indicative of moisture stress and high 
A 13 C of moisture availability or irrigation (Rehl 2008: S45). Middle and Late 
Bronze Age grains exhibited lower A 13 C values, which is in line with proxy records 
from Soreq Cave and Lake Van which also indicated less moisture in the 2nd 
millennium. In addition, drought-susceptible plants such as flax, garden pea, 
grapes, lentils, and free-threshing wheat were less evident or absent during the 
Middle as opposed to the Early Bronze Age. Moreover, the replacement of garden 
pea [Visum sativum), present in the Early Bronze Age, by vetch (Vicia ervila) 
during the Middle Bronze Age may relate to drier conditions (Rehl 2008: S49). 

These analyses provide insights into how human communities may have 
responded to a drier climate in the late 3rd and early 2nd millennium BC, but 

Introduction to Geography, Climate, Topography, and Hydrology 19 

they do not supply evidence of dramatic societal collapse. This is because we do 
not know about farming practices during phases of non-occupation, which are 
unaccounted for in these data. The changing crop types probably reflect coping 
strategies that were adopted by the population, which apparently included the 
adoption of more drought-tolerant plants during the 2nd millennium BC, as 
well as increased use of barley as fodder. As a result of these changes and a shift 
toward animal husbandry, parts of the population could continue to farm, whereas 
others adopted increasingly pastoral lifestyles (Wilkinson, Christiansen, Ur et al. 
2007: 66). 

6 Sea- Level Rise 

One of the most dramatic changes in the ancient Near Eastern environment was 
the rise in sea-level caused by the melting of polar ice caps after 18,000 BP, which 
flooded the fringing continental shelf. Because the Middle Eastern seas form part 
of the global ocean system, the rise in sea-level was roughly synchronous through- 
out the region. However, being partly cut off from the oceans, the Caspian and 
Black Seas were independent and rose and fell to the tune of local river-flow, 
rainfall, and evaporation. The Caspian continues to rise and fall dramatically and 
over the last millennium has swept inland more than 12 kilometers to obscure 
the Sasanian Gorgan Wall (Nokandeh et al. 2006; Omrani Rekavandi et al. 2007, 
2008; Sauer et al. 2009). In contrast, the dramatic flooding of the Black Sea 
basin by the rising Mediterranean c.7150 BP has resulted in its own flood myth 
(Ryan et al. 1997; but see Alcsu et al. 2002). 

Another possible origin for the flood myth derives from the flooding of what 
is now the Persian Gulf between Arabia and Iran. During the Late Glacial 
Maximum (21,000-18,000 BP), when sea-level was c. 120-130 meters below the 
present level, the entire floor of the Gulf was dry. The plain formed by the exten- 
sion of the Tigris and Euphrates rivers was dotted with occasional lakes and 
marshes, together with widespread desert extending some 1,000 kilometers from 
southern Mesopotamia to the Straits of Hormuz (Vita Finzi 1978: 258; Wilkin- 
son 2003a: 22-3; Kennett and Kennett 2007). The river draining the present 
floor of the Gulf, dubbed by some the "Ur-Schatt Rver" (Kennett and Kennett 
2007: 233-5), received tributaries from the Zagros to the north and the Arabian 
interior to the south to form a vast extended Tigris-Euphrates basin that incor- 
porated much of the present Middle East (Wilkinson 2003a: Fig. 2.3). This 
flooded gulf is important to the settlement history of the region because not only 
would it have been the locus of long-term human activity during the Late Pleis- 
tocene, but any inhabitants would have been rudely evicted by the rapid trans- 
gression that swept northwards up the Gulf in the Early Holocene into neighboring 
parts of Iran, Arabia, and Mesopotamia, perhaps contributing to founder popula- 
tions there. 

20 The Framework 

The Holocene transgression was at times rapid, especially in areas where the 
terrain was very gentle. In such locations it translated into horizontal shoreline 
movements of up to 1 kilometer per year (c. 10,000 BP) with a long-term average 
of 140 meters per year (Teller et al. 2000: 303; Kennett and Kennett 2007: 
235-6). To deal with the challenges of rapidly shifting marine and land-based 
resources, coastal communities would have needed to be mobile and adaptable. 
Needless to say, the dramatic pace of the transgression in the Gulf has encour- 
aged claims for this being the original Noah's flood (Teller et al. 2000; Kennett 
and Kennett 2007). 

Of particular significance in the development of early states is the location of 
the head of the Persian Gulf. Investigations of de Morgan (1900b), Larsen and 
Evans (1978), Sanlaville (1989) and Aqrawi (2001) suggested that during the 
5th-4th millennia BC this was located near modern Nasiriyah. Since then, a 
Belgian group has shown that in the lower Khuzestan plains of Iran the coastline 
transgressed rapidly across the shelf beneath during the initial Early Holocene 
rise of sea-level, forming a low-energy, tidal embayment by the Early and Middle 
Holocene (Baeteman et al. 2004; Gasche 2005). Following the stabilization of 
sea-level after c. 5500 BP and under more arid conditions, alluvial and coastal 
sabkhas (salt flats) started to extend and aggrade. When the sea-level rise deceler- 
ated, riverine alluviation extended the coast to the south from c. 2500 BP to its 
current position (Heyvaert and Baeteman 2007). Because Heyvaert and Baete- 
man found no evidence to support a sea-level as high as 1-2 meters above present 
levels, as suggested by Sanlaville (1989), it appears that the sea did not create a 
large marine gulf extending inland from Nasiriyah. Rather, the various borehole 
and geomorphological records suggest that when marine conditions penetrated 
c.200 milometers inland toward Nasiriyah and Amarah in southern Iraq (or 80 
kilometers in Khuzestan) around 4000 BC, lower Mesopotamia probably formed 
a complex mosaic of narrow estuaries, marsh, lagoons, and intertidal flats (both 
brackish and fresh), as suggested by Adams (1981: 16; cf. Sanlaville 1989; Aqrawi 
2001; Heyvaert and Baeteman 2007; for further discussion of how this ill-defined 
limit actually related to the economies of the cities of Ur, Telloh, and al-Hiba, 
cf. Potts 1997a and Pournelle 2007). 

In the Red Sea, the Early Holocene transgression flooded old land surfaces 
and associated shell middens, whereas off the coasts of Israel/Palestine and 
Lebanon numerous Neolithic sites were submerged. Recent off-shore investiga- 
tions have demonstrated the extraordinary potential of these submerged land- 
scapes (Galili et al. 1993; Bailey et al. 2007). 

Not only was the Early Holocene flooding of the continental shelf a significant 
event for the inhabitants of the time, it has been argued that, combined with 
increased aridity, sea-level fluctuations contributed to a complex range of human 
responses which themselves led to early state development (Kennett and Kennett 
2007: 248). Although it is difficult to separate cause and response mechanisms 

Introduction to Geography, Climate, Topography, and Hydrology 21 

in such complex cases, it is clear that sea-level rise was a major factor in the 
development of prehistoric communities. 

7 The Significance of Wetlands 

In contrast to the present landscape, the prehistoric Near East appears to have 
included many enclaves of remarkably verdancy, including wetlands. Despite the 
lack of preservation of true wetland sites, sites with some degree of organic pres- 
ervation or in the proximity of marshy areas are being increasingly recognized. 
In addition to the remarkable submarine Neolithic sites off the coast of Israel 
(Galili et al. 1993), Catal Hoyuk, Tell Oueli, and prehistoric sites in the Amuq 
plain (southern Turkey) all developed in proximity to marshes. To what degree 
wetlands were a significant factor in the formation of prehistoric communities 
and early states is unclear, but they were probably particularly significant in areas 
around the head of the Gulf near the zone inundated by the Early Holocene 
sea-level rise (Pournelle 2007). 

Wetlands appear to have played a significant role in the development of urban 
and later societies in Sumer (southern Iraq). This was not simply by contributing 
to economic sustenance, but also through the relationship of wetlands and river 
channels to belief systems and religious practice. Marshlands and the network of 
anastomosing channels were often associated with specific deities, while the 
Euphrates itself was worshipped, with particular devotion to the river gods 
attested between Mari and Hit, e.g. around Sippar, west of Baghdad, where the 
ancient course of the Euphrates bifurcated and was prone to abrupt channel 
changes or avulsions (Woods 2005; Heyvaert and Baeteman 2008). However, 
Late Holocene desiccation has not only reduced the evidence of wetlands, it has 
made it easier to diminish their significance in the development of early 

Wetlands continued to be of significance as late as Sasanian and Islamic times 
and in Mesopotamia marshes were formed as a result of the activities of Babylo- 
nian kings as well as the discharge of excess water from canals or their ultimate 
breakdown (Cole 1994; Adams 1981; Eger 2008). 

8 Geoarchaeology, Erosion, and Settlement 

Although there is a long history of investigations that have contextualized Middle 
Eastern societies within their geological environment, it was not until the 1950s 
and 1960s that Butzer (1958), Vita-Finzi (1969), and others developed models 
for the genesis of alluvial fills. Late Quaternary cycles of stream aggradation and 
incision not only provide the context for many communities in the rain-fed zone, 

22 The Framework 

they also reflect the combined effect of humans and climate. However, region- 
wide chronologies are rare (Cordova 2008: 443). Whereas Butzer saw a fine- 
grained temporal record and more complex interrelations between humans and 
the environment, Vita-Finzi focused on region-wide climatic change as a driv- 
ing factor behind fill development. Rosen (1986), Bruckner (1986), Goldberg 
(1998), and Cordova (2000) have taken an approach similar to that of Butzer, 
and in recent studies have focused upon the complexity of the relationship 
between humans and the environment as well as the role of humans in mediating 
some of the affects of climatic change (Rosen 2007). Not only are geoarchaeo- 
logical studies important for understanding the history of soil erosion, land 
degradation also provides the context for understanding the sustainability of 
ancient settlements (Cordova 2007). 

Alluvial (of rivers and wadis) and colluvial (slope) fills are complex and spatially 
variable. Broad lowland basins do not accumulate sediments uniformly; rather, 
the rate of accumulation varies from place to place so that areas around the basin 
perimeter may accumulate deep sediments, whereas those distant from sediment 
sources may receive little deposition. Hence, pessimistic statements about the 
depth of alluvial accumulation in areas such as the Marv Dasht plain of Iran 
should not be extrapolated across the entire region (Brookes et al 1982). Rather, 
as in the case of the Amuq plain (southern Turkey), sediments accumulated as a 
patchwork of alluvial fans, lacustrine deposits, alluvial fills, and anthropogenic 
sediments, all accumulating at different rates, or, in some cases, not at all (Casana 
and Wilkinson 2005). In addition, basins such as that of the Kur Rver (Iran) 
have received significant input in the form of aeolian loess in addition to alluvial 
and lacustrine deposits, with the result that prehistoric sites are not deeply buried 
(Kehl et al. 2009). The message again is that local sequences are extremely 

Although the notion that "alluvial sequences are environmentally driven but 
culturally blurred" (Macklin and Lewin 1993) also holds for many parts of the 
Middle East, it is often difficult to tell whether it was humans or environmental 
events that were the primary contributor to alluvial fills. Many geoarchaeological 
studies, understandably, emphasize the geomorphological evidence and under- 
represent the settlement record. Moreover, a focus on valley floor records at the 
expense of the valley slopes that supplied the sediments for alluvial fills means 
that such studies supply only a partial record of the sediment contribution area 
(Beach and Luzzadder-Beach 2008). However, the analysis of geoarchaeological 
records in tandem with quantitative studies of the amount and distribution of 
settlement provides a clearer idea of cause and effect relationships between human 
settlement and associated physical responses at the local level. Thus, in the Levant 
and northern Syria the significant dispersal of settlements across the landscape, 
onto slopes and some uplands, that occurred during the 1st millennium BC, and 
especially during the Hellenistic to Early Islamic periods, was associated with a 
loss of stabilizing vegetation, increased soil erosion, and more flashy flow condi- 

Introduction to Geography, Climate, Topography, and Hydrology 23 

tions, which themselves contributed to alluvial fills (Wilkinson 2003a: 146-50; 
Hill, J.B. 2004; Casana 2008). Nevertheless, vegetation loss and human distur- 
bance associated with such settlement phases simply provide the preconditions 
for soil erosion and mass movements. If the number and intensity of storms 
increase, soil erosion will exceed the capacity of the flow to remove the sediment, 
thereby resulting in an accumulation of alluvial fills. 

Clearly, both human factors and climate are contributory factors to the devel- 
opment of alluvial fills, but their relative contribution depends on local circum- 
stances. For example, in northern Jordan the history of soil development and 
climate are important for understanding long-term landscape degradation (Lucke 
et al. 2005: 65), whereas in the highlands of southwest Arabia, increased human 
setdement during a relatively dry climate initiated phases of soil erosion that fol- 
lowed long-term, stable soil development in a relatively moist climate (Wilkinson 
2005). Both human agency and climatic events are increasingly viewed as the 
combined force that led to the present, degraded landscape of the Near East. 

9 Humans and Environmental Change 

A perennial question of history and geography is: how are humans influenced by 
the environment, or alternatively, to what degree have humans had a significant 
impact on the environment? The Near East has become one focus of this debate, 
particularly with reference to societal collapse resulting from drought that led to 
famine and ultimately political or demographic collapse. In addition to cycles of 
dynastic rise and fall, certain periods, including the collapse of Akkadian settle- 
ment in northern Syria and collapse in the eastern Mediterranean c.1200 BC, 
continue to be seen, by some, as driven essentially by climatic fluctuations (Weiss 
et al. 1993; deMenocal 2001). 

Although it is clear that the environment provides opportunities and con- 
straints for agricultural productivity, equally, changes in settlement do not neces- 
sarily fall in lockstep with environmental proxy records. One well-known 
fluctuation, the 8200 BP event, which triggered a cooler, more arid climate, 
illustrates this conundrum. One group argues that this event caused major disrup- 
tions of Neolithic cultures in the Levant, northern Syria, and the eastern Mediter- 
ranean, and perhaps triggered the spread of early farmers out of the Near East 
into Greece and Bulgaria (Weninger et al. 2006). On the other hand, a Dutch 
team infers a number of cultural responses to the stresses imposed by a phase of 
cooler, arid climate. At Neolithic Tell Sabi Abyad in northern Syria, which was 
occupied throughout the 8200 BP event (dated to c.6225 BC), Neolithic life 
continued despite the fact that the site was in a climatically marginal location 
(Akkermans et al. 2010). Although cultural continuity was manifest at the site 
throughout the Neolithic, during the 8200 event several cultural changes 
occurred, including the replacement of pig husbandry by cattle, increased use of 

24 The Framework 

spindle whorls indicative of textile production, and a decline in the circulation 
of stone axes. The Dutch team argues that not only were such changes synchro- 
nous with the cool, dry event, but the climate had an indirect rather than a direct 
effect on material culture change (Akkermans et al. 2010). The implication of 
this study, supported by more than 100 radiocarbon determinations, is that 
stresses imposed by climatic events impelled communities to make changes to 
their way of life, but that these communities were resilient and successfully made 
appropriate adjustments. 

Nevertheless, in some cases the buffering mechanisms that enabled communi- 
ties to cope with shortfalls in food may have been removed as a result of climatic 
change, thereby malting communities more vulnerable to such events. For 
example, in Palestine at the end of the 3rd millennium BC the environmental 
crisis took the form of river incision that left flood plains high and dry. The 
resultant loss of potential for floodwater farming removed a buffering mechanism 
that had enabled previous communities to survive particularly dry years (Rosen 
2007: 177; Cordova 2008). 

Alternatively, in a similar environment in northern Syria, rain-fed farming was 
sustainable through most dry years if biennial fallow was practiced. However, if 
fallow was violated and cropping was annual (perhaps to sustain increasing popu- 
lations or growing towns), soil moisture was reduced, thereby jeopardizing the 
annual cereal crops (Wilkinson 1994). In addition, the drier environment of the 
late 3rd and early 2nd millennia BC may have encouraged a shift from wheat to 
barley, increased pastoral nomadism or, in parts of southern Syria, irrigation 
(Riehl 2008; Braemer et al. 2009). In addition, during the early 2nd millennium 
BC, Mari letters indicate competition over water between upstream and down- 
stream settlements in the Balikh Valley (Dossin 1974). 

By the use of multiple proxy records, together with measurements of carbon 
isotope analysis of grains from the sites of Ebla and Qatna in Syria, Roberts 
et al. (2011) have developed a climatic record that can be related to cultural 
sequences. Their record suggests that the earlier Holocene moist period was fol- 
lowed by a three-millennium-long climatic transition punctuated by three drier 
stages: at the end of the 4th millennium BC, the end of the 3rd millennium BC, 
and between 1200 and 850 BC (Roberts et al. 2011: 151). Although these appear 
to relate to phases of settlement decline, the authors also point out that the 
consequences of climatic change varied from region to region depending upon 
the sensitivities of those communities to drought (Roberts et al. 2011: 152). 
Such local variation is evident archaeologically at climatically marginal sites like, 
for example, Tell es-Sweyhat and Tell Mozan in northern Syria, both of which 
continued as significant settlements through the later part of the 3rd millennium 
BC (Danti and Zettler 2007; Pfalzner 2010). 

Over the last 3,000 years, when much of the Near East was incorporated into 
extensive territorial empires, the developing infrastructures were capable of pro- 
viding a buffer of support in times of need (Rosen 2007: 171). This is illustrated 

Introduction to Geography, Climate, Topography, and Hydrology 25 

by the colonization by Middle and Neo-Assyrian kings of the Syrian steppe, 
which, although mainly sustained by rain-fed cultivation, was associated with the 
spread of irrigation systems which became more common during the 1st millen- 
nium BC and AD (Wilkinson 2003a). Although the adoption of irrigation could 
be seen as an adaptation to a drier climate, in this case it was a result of imperial 
policy, either to make the desert bloom or to garner greater tax revenues. 

A compelling example of the advantages of expanded infrastructure comes 
from Antioch (Syria) where, in 362 AD, a poor grain harvest coincided with a 
huge build-up of troops to fight the Sasanian army. The resultant increase in 
grain prices and famine was eventually mitigated by the Roman emperor Julian 
(360-363 AD), who ordered grain imports from Chalcis and Hierapolis near 
Aleppo. If these towns had been under different ldngs, such a solution would 
have required negotiation or may have been impossible. Significantly, the grain 
was imported from the arid Membij region to verdant Antioch, an expedient 
made possible by the construction of numerous water-supply systems by previous 
Hellenistic and Roman administrations (Wilkinson et al. 2007). 

It should be emphasized that disruptions in human sustenance do not result 
simply from runs of dry years. Other factors include rainfall fluctuation, cold years 
(that kill livestock), dust storms, locust plagues, mega-floods, and river avulsions 
(Widell 2007). Not only did coping strategies vary from polity to polity, some 
- such as the huge irrigation systems of the Sasanians - may have been vulnerable 
to erosion, floods, and lack of maintenance, so that if they were severed, the 
construction required may have been too much for later communities to manage. 

10 Conclusions 

By the Roman/Arsacid period much of the landscape of the Near East was a 
product of human action. This was not simply a result of the destruction of forests 
and associated soil erosion. Marshes were formed by the discharge of excess canal 
flow into flood basins and river flow was depleted by the withdrawal of water 
for irrigation (Wilkinson 1998a). Not only do pollen records identify a spike in 
orchard development from the Levant to western Iran during the late 1st mil- 
lennium BC and early 1st millennium AD, Sumerian texts describe the range of 
trees introduced into southern Mesopotamia. These texts demonstrate that 
human management not only intensified the production of trees native to the 
area, it also brought about changes in species composition as well. 

Although the history of the Near East was one of resource depletion, pollen 
sequences demonstrate that vegetation denudation varied geographically depend- 
ing upon population levels and the scale of fuel -intensive industries such as 
mining (Barker 2000). It is now evident that increased social complexity, larger 
cities, and increased populations developed in the face of a drier, Late Holocene 
climate. Although, in general, technological and organizational developments 

26 The Framework 

enabled Near Eastern civilizations to outpace such environmental challenges, 
populations did adapt, move, and even change their lifestyle, sometimes under 
or sometimes despite the human agents or kings who led them. 


For the physical geography of the Near East, see Fisher (1978) and Sanlaville (2000). A 
more specialized, but archaeologically relevant, work on Syria is Wirth (1971). A broad 
account of the archaeological landscape and environments of the Near East is provided 
by Wilkinson (2003a). Although none of the following books is dedicated to the environ- 
ment in its entirety, Adams (1981), Potts (1997a), and Algaze (2008) describe and discuss 
the significance of the environment for Mesopotamian civilization. Van Zeist and Bottema 
(1991) remains a fundamental work on the vegetation history of the region. Different 
approaches to the eastern Mediterranean environments are provided by Bottema et al. 
(1990) and Grove and Rackham (2001). The debate concerning Early Bronze Age col- 
lapse is particularly well captured in Kuzucuoglu and Marro (2007) as well as Rosen 
(2007). An earlier collection of essays on the subject is Dalfes et al. (1997). See also 
Zettler (2003). For the Quaternary period in Saudi Arabia, see Al-Sayari and Zotl (1978) 
and some useful insights on the environment of the Arabian peninsula can be found in 
Petraglia and Rose (2009). Finally, among the large number of papers on Quaternary 
and environmental topics, the following provide useful reviews of broader interest: 
Miller (1998), Roberts et al. (2001), Brooks (2006), Robinson et al. (2006), Kennett 
and Kennett (2007), Kuzucuoglu (2007), and the papers in the special edition of the 
journal The Holocene 21/ T (2011).